Wearable Recognition Using Artificial Intelligence

Abstract

Image processing is used to extract important data from an image through processing and by using that information, machines can provide the understanding, interpretation, and description of the scene. Classifier allows us to process the image and map it to the provided classes. Segmentation is a technique to generate a black and white image of an input image based on the ground truth provided to model during the training process. Our project can do segmentation and classification on the image of cloth and use that prediction in an android application. Our application allows the user to input an image, then the segmentation model will generate a mask image and classifier identifies the wearables from that mask using convolutional neural network and classifies that wearable in provided classes. Classifiers use Deep Learning models for identification of wearables using their attributes set, the attribute set describes certain qualities about the wearable. So, by using the famous Deep Fashion Dataset, we are building a segmentation model which apply the convolutional operation of up-sampling and down-sampling on the image and generate the mask of the Region Of Interest based on the ground truth. Also, we are building a classifier/identifier to categorize the segmented images of wearables. After this, Deep Learning models is used in building a Virtual Wardrobe by categorizing the detected wearable in certain categories to facilitate the user in choosing outfits. At the end, a Virtual Wardrobe application which use these Deep Learning Model to automate the Cloth management functionality.